[GitHub] [tvm] krishnaraj36 commented on a diff in pull request #13649: [CLML][RELAY] Enable Pad and Conv2d layer fusion

[GitBox <gi...@apache.org>]

krishnaraj36 commented on code in PR #13649:
URL: https://github.com/apache/tvm/pull/13649#discussion_r1057075338


##########
python/tvm/relay/op/contrib/clml.py:
##########
@@ -200,9 +218,18 @@ def check_conv(extract):
 
         while call.op.name != "nn.conv2d":
             call = call.args[0]
+
         attrs, args = call.attrs, call.args
         if attrs.data_layout != "NCHW":
             return False
+
+        if (
+            (not isinstance(args[0], (Var, Constant)))
+            and (args[0].op.name == "nn.pad")
+            and (len(args[0].attrs["pad_width"]) != 4)

Review Comment:
   CLML module is expected to have input to 4D tensor shape for Pad layer, since conv2d layer input is always 4D tensor so Pad+conv2d fusion input is expected have 4D tensor always. It seems condition is redundant. Removed in updated commit.
   @srkreddy1238 @echuraev 



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org